Gostev wrote: ↑Jan 10, 2019 10:31 pm
Hello!
1. NDMP to Tape is strictly volume level NAS backup and restore functionality, nothing more than that. This is separate from scalable file-level backup functionality to disk (including, but not limited to NAS file shares as a source) that is scheduled to be shipped in the next update. We found that larger customers want both, they consider NDMP to Tape dumps as the "last resort" backup for when everything else fails.
2. No reasons from Veeam standpoint. As long as you can afford it (as some leading object storage providers are by an order of magnitude more expensive than tape), then this functionality can certainly be a tape replacement for long-term data archival use case.
Thanks!
Afternoon Gostev!
Will NDMP to Tape ever have a file system index/file versioning to restore an individual file via NDMP? Or is what you are saying here is that file-level backup functionality to disk will only be released on the next update (Perhaps similar to how file-to-tape works except to disk? Or Will NAS backups generate a browseable/indexable .VBK? )
We use NetApp as Filers and most file restores are covered by Previous Versions / Snap History policy (Right Click -> Properties --> Previous Versions). But when users need to go back further than the snapshot policy that is on NetApp, or for Legal Holds, we still depend on emc Networker for the NDMP client backup directly to tape which we are able to restore single files or folders via the NDMP protocol at the volume level. We also have specific application volumes that have the Snapshot policies disabled because the backups\retention policies are covered by the Networker.
In addition to all of that our NetApps are also media servers which control tape libraries over FC which bypass the production network, so I am a little worried if we switch to Veeam NDMP to tape we will be required to pull all of this data over the network to a Windows Tape server. And in the case with the coming NAS to Disk backups, targeting the paths via the \\nas\share (instead of /ndmp/volume) we would be potentially backing up terabytes & millions of files and am wondering how the file system would be indexed and how performance would be impacted on our existing topologies.
Am looking forward to someday being able to get rid of some of our legacy backup workloads!
Thanks for any information in advance
-Andy